Two experiments are reported that evaluated route learning of youth with DS, youth with Intellectual Disability and not DS (ID), and typically developing children (TD) matched on Mental Age (MA). In both experiments, participants learned routes with eight choice-points presented via computer. Several objects were placed along the route that could be used as landmarks. Participants were shown the correct route once and then were asked retraced the route without assistance. In Experiment 1 we found that the TD children and (...) ID participants performed very similarly. They learned the route in the same number of attempts, committed the same number of errors while learning the route, and recalled approximately the same number of landmarks. The participants with DS performed significantly worse on both measures of navigation (attempts and errors) and also recalled significantly fewer landmarks. In Experiment 2, we attempted to reduce TD and ID vs DS differences by focusing participants’ attention on the landmarks. Half of the participants in each group were instructed to identify the landmarks as they passed them the first time. The participants with DS again committed more errors than the participants in the ID and TD groups in the navigation task. In addition, they recalled fewer landmarks. While landmark identification improved landmark memory for both groups, it did not have a significant impact on navigation. Participants with DS still performed more poorly than did the TD and ID participants. Of additional interest, we observed that the performance of persons with DS correlated with different ability measures than did the performance of the other groups. The results the two experiments point to a problem in navigation for persons with DS that exceeds expectations based solely on intellectual level. (shrink)
This paper introduces a hybrid model that unifies connectionist, symbolic, and reinforcement learning into an integrated architecture for bottom-up skill learning in reactive sequential decision tasks. The model is designed for an agent to learn continuously from on-going experience in the world, without the use of preconceived concepts and knowledge. Both procedural skills and high-level knowledge are acquired through an agent’s experience interacting with the world. Computational experiments with the model in two domains are reported.
We present a skill learning model CLARION. Different from existing models of high-level skill learning that use a topdown approach (that is, turning declarative knowledge into procedural knowledge), we adopt a bottom-up approach toward low-level skill learning, where procedural knowledge develops first and declarative knowledge develops later. CLAR- ION is formed by integrating connectionist, reinforcement, and symbolic learning methods to perform on-line learning. We compare the model with human data in a minefield navigation task. A match between the model and (...) human data is found in several respects. (shrink)
Two experiments are reported that evaluated survey learning of youth with DS and typically developing children (TD) matched on Mental Age (MA). In Experiment 1, the experimenter navigated participants through a novel virtual environment along a circuitous path, beginning and ending at a target landmark (i.e., a door). Then, the participants were placed at a pre-specified location in the environment and instructed to navigate to the same door using the shortest possible path from their current location. They completed the task (...) three times: once after being shown the environment one time, once after three exposures, and once after five exposures. Results indicated that the participants with DS exhibited significantly less skill at identifying the shortcut than did the TD participants. In Experiment 2, participants learned two overlapping routes through a simple environment. Following acquisition, they were tested on several measures of survey knowledge: finding a shortcut, identifying the direction of landmarks not currently visible from their location in the environment, and recognizing a bird’s-eye representation of the overall environment. The results of Experiment2 indicated that the participants with DS were at least as good as the TD participants on all of our measures of survey learning. Hence, we concluded that people with DS can acquire some survey knowledge when they learn a simple environment. However, the performance of both groups was relatively poor in Experiment 2, indicating a need to consider identifying the mechanisms underlying the poor performance on wayfinding of people with DS. Identifying the underlying mechanisms, is a logical first step in devising methods of enhancing environmental learning of people with DS. (shrink)
This paper introduces a hybrid model that combines connectionist, symbolic, and reinforcement learning for tackling reactive sequential decision tasks by a situated agent. Both procedural skills and high-level symbolic representations are acquired through an agent's experience interacting with the world, in a bottom-up direction. It deals with on-line learning, that is, learning continuously from on-going experience in the world, without the use of preconstructed data sets or preconceived concepts. The model is a connectionist one based on a two-level approach proposed (...) earlier. Acknowledgements: This work is supported in part by O ce of Naval Research grant N00014-95-1-0440. (shrink)